Abstract: It is quite common in state estimation problems to have performance objectives that are expressed as the error covariance. In this paper, a sampled-data estimation covariance is defined by taking into account the intersample behavior completely. An innovation method of sampled-data estimation is also presented. The purpose of this problem is to design discrete estimators which can assign the sampieddata estimation covariance to a prespecified value such that the steady-state behavior of the state estimation is satisfactory. The sampled-data estimation problem is reduced to estimator design for a fictitious discrete-time system. Both the existence conditions and the explicit expression of the desired estimators are given. An illustrative example is provided to demonstrate the effectiveness of the design procedure.